{"id":"CVE-2026-48782","summary":"pydantic-ai: SSRF blocklist bypass via IPv4-compatible, SIIT/IVI, and local NAT64 IPv6 addresses (incomplete fix of CVE-2026-46678)","details":"Pydantic AI is a Python agent framework for building applications and workflows with Generative AI. In versions 1.56.0 through 1.101.0, 2.0.0b1, and 2.0.0b2, the cloud-metadata blocklist could be bypassed by encoding the metadata IP in an IPv6 transition form that the previous fix, CVE-2026-46678, did not decode, exposing cloud IAM short-term credentials. The previous remediation decoded only IPv4-mapped IPv6, 6to4, and the NAT64 well-known prefix, so the metadata guarantee did not hold for the remaining transition forms: IPv4-compatible IPv6 (::a.b.c.d), the NAT64 RFC 8215 local-use prefix (64:ff9b:1::/48), operator-chosen NAT64 prefixes, and ISATAP. The IPv6 wrapper is then delivered to the underlying IPv4 metadata endpoint. This occurs when an application using Pydantic AI opts a URL into force_download='allow-local' (which disables the default block on private/internal IPs) and runs on a network that actually routes the affected IPv6 transition forms: NAT64-configured networks (IPv6-only or dual-stack-with-NAT64 deployments, including some Kubernetes setups) for the NAT64 variants, or networks with an ISATAP tunnel for ISATAP. A standard dual-stack cloud VM or container does not route these forms and is not affected in practice. The IPv4-compatible and Teredo variants are deprecated and addressed as defense-in-depth. This is an incomplete fix of GHSA-cqp8-fcvh-x7r3 / CVE-2026-46678 (itself a follow-up to CVE-2026-25580). This issue has been fixed in version 2.0.0b3.","aliases":["GHSA-cg7w-rg45-pc59","PYSEC-2026-2977","PYSEC-2026-2981"],"modified":"2026-08-12T03:51:45.461006583Z","published":"2026-06-16T22:49:26.750Z","database_specific":{"cwe_ids":["CWE-918"],"osv_generated_from":"https://github.com/CVEProject/cvelistV5/tree/main/cves/2026/48xxx/CVE-2026-48782.json","cna_assigner":"GitHub_M"},"references":[{"type":"WEB","url":"https://github.com/pydantic/pydantic-ai/releases/tag/v1.102.0"},{"type":"ADVISORY","url":"https://github.com/CVEProject/cvelistV5/tree/main/cves/2026/48xxx/CVE-2026-48782.json"},{"type":"ADVISORY","url":"https://github.com/pydantic/pydantic-ai/security/advisories/GHSA-cg7w-rg45-pc59"},{"type":"ADVISORY","url":"https://nvd.nist.gov/vuln/detail/CVE-2026-48782"},{"type":"FIX","url":"https://github.com/pydantic/pydantic-ai/commit/1add06179ba4de259f7ab977620b697b7209f7e4"},{"type":"FIX","url":"https://github.com/pydantic/pydantic-ai/pull/5596"}],"affected":[{"ranges":[{"type":"GIT","repo":"https://github.com/pydantic/pydantic-ai","events":[{"introduced":"d398bc9d39aecca6530fa7486a410d5cce936301"},{"fixed":"1add06179ba4de259f7ab977620b697b7209f7e4"}],"database_specific":{"source":["CPE_RANGE","REFERENCES"],"cpe":"cpe:2.3:a:pydantic:pydantic_ai:*:*:*:*:*:python:*:*","extracted_events":[{"introduced":"1.56.0"},{"fixed":"1.102.0"}]}}],"versions":["v1.91.0","v1.101.0","v1.100.0","v1.99.0","v1.98.0","v1.97.0","v1.96.0","v1.96.1","v1.94.0","v1.95.1","v1.95.0","v1.93.0","v1.92.0","v1.90.0","v1.89.1","v1.89.0","v1.88.0","v1.87.0","v1.86.0","v1.86.1","v1.85.1","v1.85.0","v1.84.1","v1.84.0","v1.83.0","v1.82.0","v1.81.0","v1.80.0","v1.79.0","v1.78.0","v1.77.0","v1.74.0","v1.76.0","v1.75.0","v1.73.0","v1.72.0","v1.70.0","v1.69.0","v1.68.0","v1.67.0","v1.66.0","v1.65.0","v1.64.0","v1.63.0","v1.62.0","v1.61.0","v1.60.0","v1.59.0","v1.58.0","v1.57.0","v1.56.0"],"database_specific":{"source":"https://storage.googleapis.com/cve-osv-conversion/osv-output/CVE-2026-48782.json"}}],"schema_version":"1.9.0","severity":[{"type":"CVSS_V3","score":"CVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:C/C:H/I:N/A:N"}]}